Robotics Morning Review Bridge: Slide 3 / 10 / 11 v1
Date: 2026-06-13
Owner: Hugo / Genius Team
Research owner: Finance / Charlie AGT-002
Status: public-safe research artifact for morning review and Codex packaging
Visibility: PUBLIC
Target use: bridge slide / speaker notes / site intro for /robotics/why-now, /robotics/supply-chain, and /robotics/companies
Related files:
knowledge/robotics-research-harness.mdknowledge/robotics-morning-review-why-now-evidence-ladder-v1.mdknowledge/robotics-leaderdrive-morning-review-public-packaging-guardrail-v1.mdknowledge/robotics-tesla-optimus-codex-packaging-brief-v1.mdknowledge/robotics-figure-unitree-morning-review-evidence-audit-v1.mdknowledge/robotics-s5-upgrade-map-morning-review-v1.md
Public-safety: yes. No Hugo private portfolio data. No trade recommendation. No external publishing action.
0. One-line answer
截至 2026-06-13，早审最有价值的压缩方式不是再增加公司列表，而是把三个重点页面/幻灯片统一成一句话：robotics 的公开证据已经从 demo noise 走向 S3/S4 evidence tracking，但还没有跨过 S5 scaled-commercial-economics threshold。Slide 3 讲“why now: evidence unit changed”；slide 10 用 绿的谐波 / Leaderdrive 讲“filing signal is real but customer proof is missing”；slide 11 用 Tesla / Figure / Unitree 讲“三条证据曲线，不是 winner ranking”。🟠 Charlie synthesis from primary-source artifacts, as of 2026-06-13.
1. Core bridge question
Codex 明早需要解决的不是“多放几个机器人公司”，而是：
如何把 why-now、supplier case、OEM comparison 压缩成一个一致的 public-safe field guide，而不是变成交易建议、概念股宣传或公司胜负榜？
当前答案：用同一条 evidence ladder 贯穿三页。
- Slide 3: 证明 evidence unit improved。🟢/🟠
- Slide 10: 证明 supplier filing signal can be strong but still not named humanoid customer proof。🟢/🟠
- Slide 11: 证明 OEMs answer different questions: Tesla = capacity intent, Figure = deployment/manufacturing KPI, Unitree = cost-access platform。🟢/🟠
- Final guardrail: S4-heavy does not equal S5 scaled-commercial economics。🟠
2. Morning-review bridge map
| Packaging block | Public-safe title | Best evidence anchor | What it proves | What it does not prove | Signal grade | Source grade / as-of |
|---|---|---|---|---|---|---|
| Slide 3 / why-now | Why robotics now: the evidence unit changed | AI/model stack, low-cost hardware, customer-site KPI, manufacturing/capacity KPI, filing-backed revenue/supplier data | Robotics has moved beyond demo-only tracking into quantified evidence categories. | Does not prove repeatable paid deployment, ROI/payback, low intervention, uptime, margin, or durable revenue. | S3/S4 | 🟢 official sources across NVIDIA / PI / Skild / Unitree / Figure / Agility / UBTECH / Tesla / 绿的谐波; summarized 2026-06-13 |
| Slide 10 / supplier case | 绿的谐波: strong filing signal, not verified humanoid winner | 2025 revenue RMB 570.714m (+47.31%); net profit RMB 124.367m (+121.42%); harmonic reducer sales 425,158 units (+72.48%); “工业及具身智能机器人零部件” revenue RMB 422.528m, GM 34.88%; 2026-06-09 abnormal-trading announcement says 2026-06-04 to 2026-06-08 share price +36.00% vs SSE Composite -3.05% and automation-equipment index +2.13%, with no undisclosed major matter. | Filing-backed robot-relevant revenue and shipment growth are real supply-chain validation signals. | Does not prove named Tesla/Figure/Unitree/AgiBot/UBTECH customer exposure, humanoid-only revenue, contract value, single-robot value content, customer ROI, or humanoid-specific margin. | S4 | 🟢 绿的谐波 2025 annual report disclosed 2026-04-23; 2026-06-09 abnormal-trading announcement; reviewed 2026-06-13 |
| Slide 11 / OEM comparison | Tesla / Figure / Unitree: three evidence curves, not one horse race | Tesla: Fremont 1M/year designed line and Texas 10M/year long-term designed capacity. Figure: BMW 1,250+ runtime hours / 90,000+ parts loaded; BotQ 350+ Figure 03 delivered, 1/day to 1/hour cadence, >80% EOL FPY, 99.3% battery-line FPY. Unitree: R1 AIR from US$4,900 / R1 from US$5,900; G1 from US$13.5K; H2 US$29,900. | Tesla gives capacity-intent evidence; Figure gives customer-site and manufacturing KPI; Unitree gives cost-access and developer-platform evidence. | Tesla does not disclose actual output; Figure does not disclose full economics; Unitree does not disclose shipment scale, industrial uptime, support cost, or margin. | S3/S4 | 🟢 Tesla Q1 2026 Form 8-K Exhibit 99.1 filed 2026-04-22; Figure official posts dated 2025-11-19 and 2026-04-29; Unitree product pages accessed 2026-06-10; summarized 2026-06-13 |
3. Recommended bridge slide
Title: Robotics now: better evidence, not yet S5 economics
Subtitle: The field guide should track evidence types, not crown winners.
Three-card layout:
1) Why now = evidence unit changed
AI/model tools, low-cost hardware, customer-site KPI, manufacturing KPI, filings.
Signal: S3/S4 tracking quality improved.
Missing: repeat paid deployments, ROI/payback, uptime, low intervention, margin.
2) Supplier case = filing signal with guardrails
绿的谐波 has real revenue / profit / shipment evidence.
Signal: S4 supply-chain validation.
Missing: named humanoid OEM, humanoid-only revenue, contract value, customer ROI, humanoid margin.
3) OEM comparison = three curves, not winner ranking
Tesla = capacity intent; Figure = deployment/manufacturing KPI; Unitree = cost-access platform.
Signal: different S3/S4 evidence curves.
Missing: actual output, repeat orders, full economics, support cost, gross margin.
Required footer: Evidence map only. No winner ranking. No trade recommendation. S4-heavy evidence does not equal S5 scaled-commercial economics.
4. Speaker-note version for morning review
Short version:
我们不是在说“机器人已经商业化完成”。我们在说：证据单位变了。以前是 demo、TAM 和客户 logo；现在开始有 AI/toolchain、硬件价格、客户现场 KPI、制造 KPI、filing revenue 和 supplier shipments。这个变化足够让机器人从 casual watch 升级到 disciplined KPI tracking，但还不够升级成 S5 return-cycle proof。
Longer version:
Slide 3 先讲 why now：不是宏大叙事，而是证据颗粒度提升。Slide 10 用 绿的谐波做 anti-hype case：报表证据很强，但产品 fit 和收入增长不能自动推出 named humanoid OEM exposure。Slide 11 再把 Tesla / Figure / Unitree 放在三条不同证据曲线上：Tesla 的 filings 强在 capacity intent；Figure 强在 customer-site KPI 和 BotQ manufacturing KPI；Unitree 强在成本曲线和开发者可获得性。三者都重要，但都还缺 customer-confirmed economics。
5. Compression guardrails for Codex
Do say
- “Evidence unit changed” rather than “robots are solved.” 🟠
- “S3/S4-heavy evidence tracking” rather than “S5 commercialization proof.” 🟠
- “Designed capacity / long-term designed capacity” for Tesla 1M / 10M numbers. 🟢 Tesla Q1 2026 Form 8-K Exhibit 99.1.
- “Customer-site KPI and manufacturing KPI” for Figure. 🟢 Figure official posts dated 2025-11-19 and 2026-04-29.
- “Cost-access platform evidence” for Unitree. 🟢 Unitree product pages accessed 2026-06-10.
- “Filing-backed supply-chain validation” for 绿的谐波. 🟢 2025 annual report and 2026 Q1 report.
Do not say
- “Tesla is producing 1M / 10M Optimus robots per year.” 🔴 unsupported as of reviewed filings through Q1 2026.
- “Figure has proven full humanoid unit economics.” 🔴 contract value, ROI/payback, intervention rate, margin and repeat orders are not disclosed in reviewed official sources.
- “Unitree low price proves reliability, industrial adoption, shipment scale or margin.” 🔴 unsupported by official price pages.
- “绿的谐波 is confirmed supplier to Tesla / Figure / Unitree / AgiBot / UBTECH.” 🔴 named customer/order evidence not in reviewed filings.
- “Stock-price move confirms order flow.” 🔴 绿的谐波 2026-06-09 announcement says no undisclosed major matter after the +36.00% move.
- Any buy / sell / hold framing. 🔴 not allowed.
6. Signal vs noise checklist
| Claim | Treat as signal if | Treat as noise if |
|---|---|---|
| Demo video | It includes task duration, customer context, reset/failure/intervention metrics, or repeatable deployment evidence. | It only shows impressive movement without operating data. |
| Customer logo | It includes robot count, task KPI, contract value, timeline, repeat order, or ROI/payback. | It is only a pilot / LOI / partnership label without economics. |
| Capacity | It is labeled designed capacity / installed capacity / actual output correctly. | It is presented as production or shipment without output evidence. |
| Low price | It is used to explain lower experiment cost and broader access. | It is used as proof of margin, reliability, or commercial adoption. |
| Supplier revenue | It is filing-backed and tied to a clear segment/category. | It is used to infer unnamed humanoid OEM customers. |
| Share-price move | It is used only as market-context / risk-control evidence. | It is treated as order/customer evidence. |
7. What would change the morning-review conclusion
Upgrade from S4-heavy to near-S5 only if public primary/customer/filing evidence begins to show at least four of the following:
- Repeat paid deployments with named customers, robot counts, contract value and timeline. 🟢 required.
- Customer-confirmed ROI/payback or productivity improvement. 🟢 required.
- Uptime, intervention rate, safety incident rate, task success rate and service burden. 🟢 required.
- Manufacturing output, accepted deliveries, yield and cost curve. 🟢 required.
- Humanoid-specific revenue, gross margin, warranty/service cost and cash-flow contribution. 🟢 required.
- Supplier-level named customer / order / revenue / margin linkage. 🟢 required.
Downgrade if:
- Designed capacity and demo cadence improve but operating KPIs remain absent for 2-4 quarters. 🟠 monitoring rule.
- Customer logos do not convert into repeat deployments or measurable utilization. 🟠
- Low-cost hardware expands demos but support/warranty/service costs erase economics. 🟠
- Supplier revenue grows but customer concentration, ASP, margin, or architecture risk worsens. 🟢/🟠 source-dependent.
8. Public site draft section
Better evidence, not yet proof
The robotics field guide should start with a simple distinction: better evidence is not the same as completed commercialization.
By 2026, humanoid and embodied-AI robotics are no longer tracked only through viral demos. The public evidence base now includes model/tooling layers, low-cost hardware price anchors, customer-site task KPIs, manufacturing-process metrics, designed-capacity disclosures, filing-backed humanoid revenue, supplier revenue and component shipment data. That is a meaningful shift from noise toward signal.
But the sector is still not at S5. The missing evidence is economic: repeat paid deployments, customer-confirmed ROI/payback, low intervention rate, uptime, service cost, gross margin, durable revenue, and supplier customer validation. A good public page should therefore avoid ranking winners. It should teach readers how to read the evidence ladder.
9. Common misconceptions
-
“Why now means humanoids are solved.”
- Correction: why-now means evidence quality improved enough to track seriously; it does not mean return-cycle economics are proven. 🟠
-
“One strong supplier filing validates every humanoid supplier.”
- Correction: 绿的谐波 gives one strong S4 case; other suppliers still require their own customer, revenue, margin and product-fit evidence. 🟢/🟠
-
“Tesla / Figure / Unitree should be ranked as winners.”
- Correction: they answer different questions: capacity intent, deployment/manufacturing KPI, and cost-access platform evidence. 🟢/🟠
-
“A company can skip from S4 to S5 with one viral milestone.”
- Correction: S5 requires an evidence chain: paid deployment → repeatability → uptime/intervention → customer ROI → production economics → financial materiality. 🟠
10. Think Deeper questions
- Which first changes public-market valuation: customer ROI, production output, revenue disclosure, or gross margin disclosure?
- If robot bodies become cheap, does value migrate to OEMs, AI/model layers, deployment integrators, component bottlenecks, or customers?
- Should internal factory deployment count as commercialization if it produces measurable productivity/payback but no external revenue?
- What would be the cleanest supplier S5 evidence: named OEM contract, humanoid-specific revenue, margin expansion, or teardown-verified value content?
11. Source list
Primary / official source anchors:
- Tesla Q1 2026 Update, Form 8-K Exhibit 99.1, filed 2026-04-22: https://www.sec.gov/Archives/edgar/data/1318605/000162828026026551/exhibit991.htm 🟢
- Tesla Q4 2025 Update, Form 8-K Exhibit 99.1, filed 2026-01-28: https://www.sec.gov/Archives/edgar/data/1318605/000162828026003837/exhibit991.htm 🟢
- Figure, “F.02 Contributed to the Production of 30,000 Cars at BMW,” 2025-11-19: https://www.figure.ai/news/production-at-bmw 🟢
- Figure, “Ramping Figure 03 Production,” 2026-04-29: https://www.figure.ai/news/ramping-figure-03-production 🟢
- Unitree R1 product page, accessed in existing source-backed artifact 2026-06-10: https://www.unitree.com/R1 🟢
- Unitree G1 product page, accessed in existing source-backed artifact 2026-06-10: https://www.unitree.com/g1/ 🟢
- Unitree H2 product page, accessed in existing source-backed artifact 2026-06-10: https://www.unitree.com/H2 🟢
- 绿的谐波 2025 annual report, disclosed 2026-04-23: https://static.sse.com.cn/disclosure/listedinfo/announcement/c/new/2026-04-23/688017_20260423_EU45.pdf 🟢
- 绿的谐波 2026-06-09 abnormal trading announcement: https://static.sse.com.cn/disclosure/listedinfo/announcement/c/new/2026-06-09/688017_20260609_MZU2.pdf 🟢
- UBTECH Robotics Corp Ltd Annual Report 2025, disclosed 2026-04-15: https://owebsite-cdn.ubtrobot.com/resources/file/2026/04/15/795097683329093.pdf 🟢
- NVIDIA Isaac GR00T official developer page, captured in existing source-backed artifact 2026-06-09: https://developer.nvidia.com/isaac/gr00t 🟢
Internal synthesis cross-references:
knowledge/robotics-morning-review-why-now-evidence-ladder-v1.md🟠knowledge/robotics-leaderdrive-morning-review-public-packaging-guardrail-v1.md🟠knowledge/robotics-tesla-optimus-codex-packaging-brief-v1.md🟠knowledge/robotics-figure-unitree-morning-review-evidence-audit-v1.md🟠knowledge/robotics-s5-upgrade-map-morning-review-v1.md🟠
12. Public-safety flag
Public-safe if used as an evidence map and not as a recommendation.
Do not publish:
- Hugo private portfolio data.
- Buy / sell / hold language.
- Winner ranking.
- Claims that S4 signals equal S5 commercialization.
- Claims that designed capacity equals current production.
- Supplier-customer inferences without named customer, order, filing, teardown, or revenue evidence.